It's more complicated than 'science = good, testable, unbiased search for the truth'. It can be but it is more complicated than that.
A lot of stuff that appears to be medical 'science' is actually written by PR companies. There are plenty of jobs out there for 'medical writers' in PR agencies. A recent study in PLoS Medicine found that 41% of abstracts issued by researchers and 46% of press releases issued by researchers, contained ‘spin’ i.e. claims that were not supported by the evidence, such as reporting a finding that is statistically insignificant as if it were relevant.
The positive thing, of course, is that it was scientists who did this study into spin. But there are plenty of other scientists who put their names to abstracts that conveniently exaggerated the effectiveness of new medicines.
There's a rule of thumb that when a new drug is licensed, there will be lots of studies that say it is very effective, yet within 10 years the weight of evidence often shows it's little better than placebo. (Clearly this does not apply to every drug in every condition, it's a rough generalisation.)
Drug companies are not obliged to publish all studies, positive and negative. Peer review, while A Good Thing in terms of senior people with expertise evaluating research, can also be problematic, if the reviewers are committed to the consensus and react badly to a paper that threatens their assumptions, or if they don't declare conflicts of interest. Or if they just don't like the researchers or the institution- peer reviewers are human beings who can be petty, or biased, or whatever.
You would think it is in the interest of drug companies to make safe, effective drugs, but actually it can be in their short-term interest to downplay safety concerns. Look at Seroxat and suicide if you want to see the pharmaceutical industry at its worst - and not only the industry but some eminent scientists and supposedly reputable professional bodies. Admitting your drug is killing people, or doesn't work, tends not to be great for your profits or your shareholders.
Scientists are human beings. They are not infallible. But beyond that, there are some wider issues with medical research and the profit motive. I don't know enough about other sciences to judge whether similar issues apply there, but I don't see why they would be exempt from the failings of humanity.